Community perceptions and practices of treatment seeking for childhood pneumonia: a mixed methods study in a rural district, Ghana

This study was carried out in the Dangme west district of Greater Accra region in Ghana. At the time of the study, the district had an estimated land size of 1,700 square kilometres with a population of about 109 459 and 376 communities or villages [18]. Available health facilities at the time of the study included four government owned health centres, six community clinics, two privately owned clinics, two private maternity homes, two pharmacies, and 42 registered drug retail shops. The district had no hospitals. Severe cases were referred to hospitals in neighbouring districts. The indigenes are of the Ga-Adangme ethnic group and the dominant occupation is trading followed by subsistence farming.

The study employed a mixed methods design to explore community perceptions and practices in the management of childhood fevers with emphasis on respiratory symptoms (especially pneumonia).

In addition to the KI interviews, we conducted a household survey of a representative sample of 501 caregivers. Details of the sample were accessed via the database of the Dodowa Health and Demographic Surveillance System (DHDSS). This database comprises of information on socio-demographic characteristics of the district population; such as age, sex, marital status and total number of children per woman. Other variables are a unique identification code for each caregiver, community of residence and house address.

The respondents were selected through a multistage sampling technique as follows: 15 communities were randomly selected from each of the four sub districts. All houses within the selected communities with under-fives were selected, and from each selected house, only one eligible caregiver from a household was selected for interview, by simple random sampling.

We purposively recruited key Informants and FGD participants with the assistance of a community mobilization officer who lived in the community. Research team members screened recruits for eligibility and selection prior to the day of discussion. The discussions were held in quiet places within the communities and in the local language, Dangme, to allow participants to express themselves more easily and to elicit local terminology that could be useful in designing messages for the intervention. A social scientist who spoke Dangme fluently conducted the Key Informant interviews and moderated the focus groups. She was assisted by a note taker who was proficient in Dangme and the first author who only had basic knowledge of the language. A list of topics, complemented with a video on Acute Respiratory illnesses (ARI) for training health workers in Integrated Management of Childhood Illnesses (IMCI) was used to guide the discussions. Topics discussed included caregiver perceptions of common childhood illnesses in their communities, breathing difficulties among children, recognition of childhood pneumonia, the local name for pneumonia and its treatment seeking practices (Assessment of pneumonia was based on IMCI criteria i.e. the presence of cough, difficult or fast breathing with or without fever). The video sections showed a child suffering from pneumonia; with fast breathing and chest in drawing. The use of the video was to help participants see the illness symptoms and to generate more terms related to the illness from the participants [20]. The discussions were audio recorded with consent from participants. Each discussion lasted about one hour to one hour thirty minutes.

The qualitative data was reviewed to extract repeated ideas or themes that were apparent. Themes that emerged from the key Informants and focus group discussions were not only grouped into categories but were further used to design the survey instrument.

The household survey assessed caregiver perceptions, and practices in the management of fevers and breathing difficulties with emphasis on pneumonia in children under five. Information sought from the respondents included caregivers’ perceptions of common childhood illnesses in the communities, perceptions of severe childhood illnesses, recognition of the symptoms and causes of pneumonia, treatment seeking practices for the disease and the local name for pneumonia. Other variables covered demographic characteristics of the caregivers and whether or not caregivers were willing to utilize CHW services for treatment of fever in under-fives if made available and actions taken at home for childhood fevers.

A team of eight trained research assistants (Six interviewers and two serving as supervisors) conducted face to face interviews with caregivers in their homes. Interviews were held in the local languages using a questionnaire with both open and closed ended questions. Several role play sessions of translating questions from English to the local languages and field pretesting among the data collection team ensured adequate practice prior to actual field data collection. Supervisors helped to resolve any emerging challenges during the field data collection and also checked filled questionnaires for completeness. Data collection lasted four weeks for both qualitative and quantitative survey.

The household data was coded, categorized, entered in SPSS and was exported and analyzed in STATA. Descriptive and inference statistics were the primary tools used to explore and interpret findings of the quantitative survey data. To examine the factors that determine the use or otherwise of CHW services, probability (probit) estimation technique was employed for this study because it uses maximum likelihood estimation. Furthermore, the dependent variable is a binary dummy variable which takes value of 1 if CHW services are used and 0 if CHW services are not used [21]. The dependent variable in this study is the intention to use the services of the CHWs or not. The explanatory variables included the age of the caregiver, sex, religion, education, constraints (financial, unfriendly staff, and proximity to the CHW), marital status and birth parity (number of children) of the respondents.

The tape-recorded FGDs and KIIs were transcribed verbatim, supplemented with field notes and any identifying information removed to ensure participants anonymity [19]. We adopted the grounded theory approach to the data analysis. The first author and the moderator of the interviews read all transcripts several times to familiarize with the data, identify key themes and develop a coding scheme. The transcripts were analyzed and coded. We compared coded themes and discussed discrepancies until agreement was reached. We then grouped segments of the interviews under the relevant codes and further analyzed the data by sub themes.

The caregivers were predominantly female, and aged between 25 and 34 years, as shown in Table 1. Twenty nine percent of the respondents had no formal education. Most were Christians (87.6 %) and married (87.8 %).

The caregivers’ knowledge about pneumonia is shown in Table 2. The results suggest that, perceptions and understanding of illnesses that affect under-fives, and adequate knowledge about their causes, effects and treatments are context specific. Most respondents were inadequately informed about pneumonia, its signs, symptoms, causes and treatment. Only a third of respondents 31.94 % (160/501) indicated they had ever heard the name pneumonia. Of those who reported having ever heard of pneumonia, majority 58.13 % (93/160) did not know any symptoms of pneumonia as shown in Table 2. Only 5.62 % (9/160) mentioned a word in Dangme for pneumonia and each name varies.

Majority, 96.6 % affirmed their willingness to utilize the CHW services for their children if such services are made available in their communities. The probit regression showed that age, education, religion, sex and occupation are not important predictors of the intention (willingness) of caregivers to use services of the CHWs for childhood fever. Marital status and birth parity however significantly influence a caregiver’s decision to use services of the CHWs (Table 3). Married respondents were more likely than the unmarried to use services of the CHWs (P < 0.001). The use of these services also increases with an increasing parity (P < 0.001). Financial challenges and proximity to the CHW’s location are some reasons why respondents may not use the services of CHWs.

The respondents gave a clear description of various treatment practices for fever. As shown in Table 4, the child is usually given a tepid sponging (61.9 %) to bring down the temperature, followed by self-administered home treatment with remnants of orthodox medicine kept from a previous ailment or herbal remedies. The common medication administered at home were mostly analgesics such as paracetamol (92.4 %) and anti-malarial, mainly chloroquine (33.1 %). Majority, (93.4 %) would normally initiate treatment on the same day or within 24 h from when symptoms are observed.

Participants in the FGDs comprised of young mothers, older mothers, grandmothers and fathers. The young mothers were aged between 21 and 32 years and had one to three children, some were married. Many of these mothers have primary and/or Junior Secondary School education while few of them have no formal education. They had lived in the community most of their lives. Most of these younger mothers were traders or artisans. A few were unemployed.

Older mothers were aged between 35 and 47 years. Most were married while some were divorced or single parents. Almost all the older mothers had little or no formal education and had lived in the community since birth. The grandmothers were engaged in either farming or trading or both, were averaged between 50 and 77 years, and had lived in the community most of their life time.

Men who were engaged in the FGDs were between the ages of 20 and 43 years and were mostly married. All male participants were employed, and had lived in the community between 10 and 38 years. Most of them were farmers and the rest artisans. All of them had some formal education with the highest being secondary education.

Key issues explored from the various study participants in the qualitative interviews yielded similar responses therefore the findings are presented together and under three main themes: Community perception of common childhood illnesses; perceived causes of difficulty in breathing and or pneumonia and Treatment seeking for these illnesses.

On the whole, “difficulty in breathing or breathing difficulties” was rarely mentioned spontaneously as one of the common illnesses among children in the district. Further probing revealed that difficulty in breathing in children was often associated with “soso” (common cold) or “blocked nose”, “hoha” (asthma) and convulsion.

However, after FGD participants and KIs viewed the video of a child with high respiratory rate, cough and lower chest in-drawing, indicating severe pneumonia (WHO’s IMCI classification), majority of the discussants generally described the child in the video as having breathing problems. Very few (about 11 % of 56 of FGD participants) including a herbalist spontaneously identified the child as suffering from pneumonia. A male participant in one of the FGDs exclaimed “…oh that is pneumonia; it is common among adults but not the children”.

When probed further, majority said they have never heard the word “pneumonia” before, adding that it was their first time of hearing it. Others, who reported they have heard about pneumonia, indicated the radio, school and clinic as their source of information.

It appeared that there was no single word or expression used to describe “pneumonia” in the local dialect. Various expressions used in the Dangme dialect to describe the condition included “emumu-sue-no” or “emumu-siueo-nowie” (meaning “gasping for breath”) and “emumu-dei-no” or “enge-mumu woe mla mla mla” (meaning “the child is breathing faster than normal”). The expressions given by respondents in the household survey included: “… kasa hio” (meaning “illness of the rib area”); “fie se ehem” (meaning “cold air has entered into his/her body”).

Additionally, the herbalist, who identified the child as suffering from pneumonia, suggested the name for pneumonia in the local dialect as “Mumunya tam hio” (meaning, “out of breath” illness). The variety in common expressions among the participants suggest inadequate exposure/knowledge or understanding of pneumonia except for those who have ever seen someone suffer from it or someone who had been diagnosed of it as told by one female participant. She said “I went to visit my relative on admission to the hospital and she told me the Nurse told her that she had pneumonia…she was not breathing properly but I don’t know a name for it in Dangme”.

Most participants perceived causes of breathing difficulties in children as common cold, headache, cough, asthma, hot body and ‘hidden measles’ (this is when a child has other symptoms of measles without the skin rash), eating ‘Bad food,’ (junk food) and being infested with worms and measles.

The perception of caregivers who have heard about pneumonia differed. The general belief was that exposing a child, to cold temperature or windy weather conditions could result in the child getting pneumonia. The most common factors mentioned in all discussion groups included “sleeping on the bare cemented floor”. For instance, one female participant said “What I know is that pneumonia is caused by sleeping on the bare cemented floor; because the bare floor is cold, if the child is not adequately clothed and sleeps on it for a long time, the cold penetrates the body and causes pneumonia.”

Another belief expressed by a male discussant was that, “Sleeping directly under a fan can cause a child to have pneumonia; I mean the continuous exposure of a child to the air from an electric fan over a long period of time makes the child sick with pneumonia.”

Furthermore, other views sampled attributed the cause of pneumonia to ingesting cold drinks or food; they indicated that, “One gets pneumonia from drinking cold water, eating cold food or chewing ice cubes.” The Herbalist also added that, “… In the past, when there was no iced water children did not get pneumonia.”

Overall, treatment practices that were mentioned by participants can be grouped into Self-treatment and Hospital treatment.

Assessing actual behaviour of study participants would have been more ideal than exploring perceptions of participants since this may have revealed their actual practice. The use of a triangulation of quantitative and qualitative methods with different participants however strengthened our findings from the different community perspectives explored. The use of the video showing real children with pneumonia proved useful. It enriched the discussions as it elicited more information from participants compared with the use of the discussion guide alone [20]. The study findings could be largely generalized to the district due to the probability sampling method used in selecting respondents for the household survey.